@mtalhaarshad, You can checkout @abhishek youtube channel[https://www.youtube.com/c/AbhishekThakurAbhi/], he has explained this problem and many other also.
", "rawMarkdown": "mtalhaarshad, You can checkout @abhishek youtube channel[https://www.youtube.com/c/AbhishekThakurAbhi/], he has explained this problem and many other also.", "votes": null }, { "id": "916455", "postDate": "07/05/2020 17:00:20", "content": "Perhaps you should use the official topic New to Machine Learning or Kaggle? which is created for this very purpose, and avoid creating multiple new Topics in the Discussions forum unnecessarily.
", "rawMarkdown": "Perhaps you should use the official topic [New to Machine Learning or Kaggle?](https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/154297) which is created for this very purpose, and avoid creating multiple new Topics in the Discussions forum unnecessarily.", "votes": null }, { "id": "916463", "postDate": "07/05/2020 17:19:34", "content": "Everyone uses transfer learning. You can download pretrained state of the art image classifiers from the internet and then fine tune them on the competition data. I share an overview of how to approach any Kaggle image classification competition here. Specifically \"step 3\" answers your question.
", "rawMarkdown": "Everyone uses transfer learning. You can download pretrained state of the art image classifiers from the internet and then fine tune them on the competition data. I share an overview of how to approach any Kaggle image classification competition [here][1]. Specifically \"step 3\" answers your question.\n\n[1]: https://www.kaggle.com/cdeotte/how-to-compete-with-gpus-workshop", "votes": null }, { "id": "916588", "postDate": "07/05/2020 19:57:56", "content": "I would say even in university research we tend to use transfer learning. For my masters thesis I am using an ensemble of ResNet50 networks to classify abnormal frames in wireless capsule endoscopy. Designing your own architecture can be very tedious and generally leads to less than satisfactory results.
", "rawMarkdown": "I would say even in university research we tend to use transfer learning. For my masters thesis I am using an ensemble of ResNet50 networks to classify abnormal frames in wireless capsule endoscopy. Designing your own architecture can be very tedious and generally leads to less than satisfactory results.", "votes": null }, { "id": "917550", "postDate": "07/06/2020 15:54:13", "content": "How to choose which model will be good for a specific challenge? or if to concatenate different models how to choose them?
", "rawMarkdown": "How to choose which model will be good for a specific challenge? or if to concatenate different models how to choose them?", "votes": null }, { "id": "917910", "postDate": "07/06/2020 19:48:28", "content": "It really depends on the problem you are facing. Only by experimenting you will gain meaningfull insights that may help you taking that decisions.
", "rawMarkdown": "It really depends on the problem you are facing. Only by experimenting you will gain meaningfull insights that may help you taking that decisions.", "votes": null } ], "comments": [ { "id": 916443, "author_name": "orionpax00", "author_url": "", "post_date": "07/05/2020 16:45:35", "content": "@mtalhaarshad, You can checkout @abhishek youtube channel[https://www.youtube.com/c/AbhishekThakurAbhi/], he has explained this problem and many other also.
", "votes": null, "replies": [] }, { "id": 916455, "author_name": "sirishks", "author_url": "", "post_date": "07/05/2020 17:00:20", "content": "Perhaps you should use the official topic New to Machine Learning or Kaggle? which is created for this very purpose, and avoid creating multiple new Topics in the Discussions forum unnecessarily.
", "votes": null, "replies": [] }, { "id": 916463, "author_name": "cdeotte", "author_url": "", "post_date": "07/05/2020 17:19:34", "content": "Everyone uses transfer learning. You can download pretrained state of the art image classifiers from the internet and then fine tune them on the competition data. I share an overview of how to approach any Kaggle image classification competition here. Specifically \"step 3\" answers your question.
", "votes": null, "replies": [] }, { "id": 916588, "author_name": "hmorera", "author_url": "", "post_date": "07/05/2020 19:57:56", "content": "I would say even in university research we tend to use transfer learning. For my masters thesis I am using an ensemble of ResNet50 networks to classify abnormal frames in wireless capsule endoscopy. Designing your own architecture can be very tedious and generally leads to less than satisfactory results.
", "votes": null, "replies": [ { "id": 917550, "author_name": "mtalhaarshad", "author_url": "", "post_date": "07/06/2020 15:54:13", "content": "How to choose which model will be good for a specific challenge? or if to concatenate different models how to choose them?
", "votes": null, "replies": [] }, { "id": 917910, "author_name": "cayala", "author_url": "", "post_date": "07/06/2020 19:48:28", "content": "It really depends on the problem you are facing. Only by experimenting you will gain meaningfull insights that may help you taking that decisions.
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "916428": "", "916443": "mtalhaarshad, You can checkout @abhishek youtube channel[https://www.youtube.com/c/AbhishekThakurAbhi/], he has explained this problem and many other also.", "916455": "Perhaps you should use the official topic [New to Machine Learning or Kaggle?](https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/154297) which is created for this very purpose, and avoid creating multiple new Topics in the Discussions forum unnecessarily.", "916463": "Everyone uses transfer learning. You can download pretrained state of the art image classifiers from the internet and then fine tune them on the competition data. I share an overview of how to approach any Kaggle image classification competition [here][1]. Specifically \"step 3\" answers your question.\n\n[1]: https://www.kaggle.com/cdeotte/how-to-compete-with-gpus-workshop", "916588": "I would say even in university research we tend to use transfer learning. For my masters thesis I am using an ensemble of ResNet50 networks to classify abnormal frames in wireless capsule endoscopy. Designing your own architecture can be very tedious and generally leads to less than satisfactory results.", "917550": "How to choose which model will be good for a specific challenge? or if to concatenate different models how to choose them?", "917910": "It really depends on the problem you are facing. Only by experimenting you will gain meaningfull insights that may help you taking that decisions." }, "source": "meta" }